Caleb Mwangi manages an eight-person operations team at a logistics company. One Thursday afternoon a senior analyst on his team, normally unflappable, lingered after a one-on-one and asked the question Caleb had been dreading: "Be honest with me. In two years, is AI going to do my job?" Caleb did not have a clean answer, and pretending he did would have been worse than admitting he did not. So he said, "I do not know exactly what changes. But I would rather we figure out together which parts of your work are likely to shift, and get ahead of it, than wait and react." That conversation became the reason he sat down the next week and built his first scenario plan. Not for the company. For his team.

What This Lesson Covers

Scenario planning means preparing for several possible futures instead of betting on one prediction. You will not learn to forecast exactly how AI changes your team's work, because nobody can. Instead you will learn to sketch a small number of plausible futures, decide what your team should do under each, and watch for the signals that tell you which future is actually arriving.

This lesson keeps a tight scope: your team. Company-wide and industry-wide disruption strategy belongs to senior leaders. Your job is the work in front of your people: which of their daily tasks are most exposed to AI, what reskilling paths make sense, how to redeploy capacity that AI frees up, how to redesign roles, and how to talk about all of it honestly so your team feels prepared instead of afraid.

Before going further, a few plain definitions. Disruption means a change large enough that the old way of working stops being the best way. A task is one specific activity, like "summarize a shipment exception report." A job is a bundle of many tasks. AI rarely replaces a whole job at once; it changes individual tasks. Automation means AI does a task with little human involvement. Augmentation means AI helps a person do a task faster or better while the person stays in control. Keeping task and job separate, and automation and augmentation separate, is the whole game here.

You cannot predict which future arrives. You can make sure your team is ready for more than one of them.

Why Forecasting Falls Short

Most planning starts with a forecast. You predict the future and plan accordingly. The problem is that forecasts are frequently wrong. You make the prediction, the world moves somewhere else, and your plan quietly stops matching reality while you continue executing it.

Scenario planning refuses the premise. Instead of predicting one future, you explore several and prepare for more than one outcome. That is simply a more honest description of the situation you are in. The future genuinely is uncertain, multiple outcomes genuinely are possible, and a strategy that holds up across several of them is more robust than a strategy that only works if your forecast happens to be exactly right.

This is worth saying plainly because it changes what success looks like. Scenario planning is not about being right about what will happen. It is about exploring what could happen and positioning your team, and eventually your organization, to do well under more than one version of it.

Why Task-Level Thinking Beats Job-Level Panic

When Caleb's analyst asked "will AI do my job," the question itself was the trap. A job is too big a unit to reason about. The analyst's role bundles maybe a dozen distinct tasks, and they are not equally exposed to AI. Some are highly exposed: drafting routine status emails, reconciling two spreadsheets, summarizing long reports. Some are barely exposed at all: negotiating a delivery exception with a frustrated client, judging whether an unusual shipping pattern signals fraud, mentoring a new hire.

So the honest answer is almost never "yes, AI takes your job" or "no, you are safe." It is "three of your twelve tasks are likely to be mostly automated, four will be augmented so you do them faster, and five are not going anywhere soon, and here is what we do with the time that frees up." That is a far less frightening sentence, and it is also far more useful for planning.

This is why every step below works at the task level. You break the job into tasks, rate each task's exposure, and only then think about people and roles.

Step 1: Map Your Team's Task Exposure

Start by listing what your team actually spends time on. Not the org chart version, the real version. For each person or role, write down the recurring tasks and roughly what share of their week each one takes. Then rate each task on two things: how exposed it is to AI today, and whether the likely AI relationship is automation (AI does it) or augmentation (AI helps).

Use a simple scale so you can move fast. Exposure: High, Medium, Low. Relationship: Automate, Augment, or Neither. You are not aiming for precision. You are aiming for a shared picture your team can look at together.

Here is the worked inventory Caleb built for his operations team. The percentages are how much of the team's total working time goes to each task, estimated from a normal month.

• Manual data reconciliation between systems (18% of team time): Exposure High, relationship Automate. Matching records across two systems is repetitive and rule-based.
• Drafting routine status updates and reports (14%): Exposure High, relationship Augment. AI drafts; a person checks and sends.
• Summarizing long exception and incident reports (10%): Exposure High, relationship Augment.
• Answering repetitive internal questions (9%): Exposure High, relationship Automate, via an internal assistant.
• Investigating unusual shipping anomalies (12%): Exposure Medium, relationship Augment. AI flags candidates; humans judge.
• Resolving exceptions with carriers and clients (15%): Exposure Low, relationship Neither. Needs negotiation and relationship judgment.
• Coordinating across departments on escalations (12%): Exposure Low, relationship Neither.
• Coaching and onboarding teammates (10%): Exposure Low, relationship Neither.

Add up the high-exposure-to-automate-or-augment tasks and a pattern jumps out. Roughly 51% of the team's time sits in High-exposure work. That is not a layoff signal. It is a redeployment signal: a large chunk of hours is about to get cheaper to produce, and the question is what those hours become.

You can use AI to build this inventory faster. Caleb pasted his task list into an AI assistant with this prompt: "Here is a list of tasks my operations team performs, with rough time percentages. For each task, rate AI exposure as High, Medium, or Low, and say whether the likely relationship is automation or augmentation. Explain each rating in one sentence. Then flag any task where you are uncertain." The draft was a useful starting point. Caleb overrode two ratings the AI got wrong, because it did not know that carrier negotiations at his company involve handshake relationships built over years. The AI gives you a first cut; your knowledge of the team corrects it.

Step 2: Build Three Plausible Futures

You do not need a fancy 2x2 matrix for a team-level plan. Three clearly different futures are enough to stress-test your thinking. The point of writing more than one is to stop yourself from quietly assuming the future you find most comfortable.

Caleb built these three, framed entirely around what they mean for his eight people.

Future A, Gradual Augmentation. AI tools improve steadily but not explosively. The team adopts them task by task. Most high-exposure tasks become augmented rather than fully automated; people stay in the loop and simply move faster. Headcount stays flat, but each person handles more volume and spends a larger share of time on judgment work. This is the most likely future and the easiest to manage.

Future B, Fast Automation. AI capability jumps quickly. Reconciliation, routine drafting, and repetitive question-answering become almost fully automated within a year. A meaningful slice of the team's current work simply disappears as a human activity. The risk here is not just job loss; it is that the team's identity is tied to tasks that vanish. The opportunity is that freed capacity is large enough to take on work the team never had time for.

Future C, Regulation Slows It Down. Privacy rules, data-handling restrictions, or internal compliance limits mean AI can only be used in narrow, approved ways. Many exposed tasks stay manual longer than expected because the tooling cannot be cleared for use. In this future the premium shifts to people who understand the rules and can run AI safely within them. Augmentation happens, but slowly and under supervision.

For each future, Caleb answered four questions in writing: What changes for my team? What new opportunities open up? What is the biggest risk to my people? What would we need to start doing now to be ready? Writing the answers, even roughly, is what turns scenarios from an interesting exercise into a plan.

Step 3: Project the Freed Hours Under Each Future

Here is where the inventory pays off. Caleb's team of eight works roughly 8 people times 40 hours, about 320 hours a week. Take the high-exposure tasks and estimate how much of that time AI could free under each future. Freed does not mean lost; it means available to redirect.

The High-exposure tasks (reconciliation 18%, routine drafting 14%, report summarizing 10%, repetitive questions 9%) total 51% of time, which is about 163 hours a week.

• Future A, Gradual Augmentation: assume AI saves about 30% of the time spent on those high-exposure tasks. That frees roughly 49 hours a week, a little over one full person's worth of capacity.
• Future B, Fast Automation: assume AI removes about 65% of that time. That frees roughly 106 hours a week, close to two and a half people's worth of capacity.
• Future C, Regulation Slows It: assume AI is cleared for only part of the work and saves about 12%. That frees roughly 20 hours a week, half a person's worth.

These numbers are illustrative, built from Caleb's own estimates, not from any outside study. But they make the strategic question concrete. In Future A he has about one person's worth of hours to reinvest. In Future B he has nearly two and a half, which is large enough that doing nothing would look like overstaffing. The plan he needs is the same in shape across all three: decide in advance where freed hours go, so the team grows into higher-value work instead of being seen as surplus.

Step 4: Plan the Team's Response

Freed capacity is only good news if you have already decided what it becomes. Caleb planned three responses, because the freed hours have to go somewhere useful.

Redeploy freed capacity toward low-exposure, high-value work. The team's Low-exposure tasks, carrier negotiation, cross-department escalation, anomaly investigation, are exactly the work that is always under-resourced because the routine work crowds it out. Caleb's plan: as reconciliation and drafting hours shrink, route that time into deeper anomaly investigation and proactive client relationship work. The team does not get smaller; it gets more valuable per hour.

Build reskilling paths. Reskilling means deliberately helping people build skills for where the work is heading. Caleb mapped a path for each role. For the analyst whose reconciliation work was most exposed, the path was: learn to supervise and audit AI outputs, then deepen the anomaly-investigation judgment that AI cannot do. He set a concrete first step, two hours a week of protected learning time, and a check-in at 60 days. Reskilling fails when it is a vague encouragement to "upskill." It works when it is a named skill, a scheduled time, and a follow-up.

Redesign roles around what is left. If half a role's tasks get automated, the role should not be left as a hollowed-out version of itself. Redesign it. Caleb sketched a future version of the analyst role that was 60% judgment and relationship work, 30% supervising AI-augmented workflows, and 10% routine, instead of today's near-reverse split. Sharing that redesigned picture early gave his analyst something to move toward rather than something to fear.

You can use AI to draft a skill-gap analysis here. Caleb's prompt: "Here is a current role description and a redesigned future version of the same role. List the specific skills the person would need to build to move from the current version to the future one, ordered from most to least urgent, and suggest a realistic way to build each one." He treated the output as a first draft and adjusted it with what he knew about the individual.

Step 5: Spot the Early Signals

You will not get an announcement telling you which future arrived. You watch for leading indicators, small early signs that one future is becoming more likely than the others, and you adjust before you are forced to.

For each of Caleb's three futures, he wrote down what he would watch for:

• Signals of Future B, Fast Automation: a reconciliation tool the team pilots suddenly handles edge cases it used to fail on; the vendor ships major capability jumps every few weeks; other teams report tasks fully handed over to AI. If Caleb sees these, he accelerates reskilling and pulls the role redesign forward.
• Signals of Future C, Regulation Slows It: the company's legal or security team starts reviewing AI tools more strictly; an approved tool gets pulled pending review; new data-handling rules land. If Caleb sees these, he shifts investment toward people who can run AI compliantly and slows aggressive automation plans.
• Signals of Future A, Gradual Augmentation: steady but unspectacular improvement, adoption happening task by task, no sudden cliffs. This is the default he assumes until a stronger signal points elsewhere.

The discipline is to name the signals before they appear. It is much easier to spot a trend when you decided in advance what to look for than to notice it while you are busy.

Step 6: Communicate Honestly to Reduce Fear

This is the part managers most often get wrong, in one of two directions. Some go silent, hoping not to alarm anyone, which leaves people to imagine the worst. Others over-reassure with "your job is totally safe," which is a promise you cannot keep and which destroys trust the moment it proves false.

The honest middle is what worked for Caleb. He brought the task inventory to the team and showed them the real picture: here are the tasks likely to be automated, here are the ones likely to be augmented, here are the ones that are not going anywhere, and here is the plan for the freed time. He was specific about what he did not know, and specific about what he was committing to: protected learning time, a redesigned role to grow into, and a promise to redeploy freed capacity rather than treat people as surplus.

A few principles made the conversation land:

• Name the change, do not hide it. People can feel disruption coming. Acknowledging it openly is less frightening than a manager pretending nothing is happening.
• Separate task from job out loud. Saying "AI will take over reconciliation, not your role" is both more accurate and far less scary than leaving "AI is coming" to hang in the air.
• Give people agency. A reskilling path turns a passive fear into an active project. People who are building toward something specific worry less.
• Do not promise what you cannot guarantee. Commit to the things you control, support, time, honest updates, rather than to outcomes you do not.

Step 7: Revisit the Plan on a Schedule

A scenario plan built once and filed away is worse than useless, because it gives false confidence while quietly going stale. The world moves, uncertainties resolve, and last quarter's futures stop matching reality.

Caleb put a recurring 45-minute review on his calendar every quarter. In it he asks four questions: Which of my early signals have I actually seen? Has any future become clearly more or less likely? Is the task inventory still accurate, or have tasks shifted? Are the reskilling paths on track, and do they still point the right way? He updates the plan, communicates any change to the team, and moves on. The review is short on purpose. The goal is a living plan, not a perfect document.

This guards against the most common failure mode: not panic, not paralysis, but neglect. The plan that protects your team is the one you keep current.

Widening the Lens: Six Ways AI Disruption Could Play Out

Caleb's three futures were built around his eight people. Sooner or later the same discipline has to be applied at a wider altitude, because the forces that reshape your team's tasks originate outside your department. When you are asked to contribute to that conversation, or when you simply want to understand the weather your team is operating in, it helps to have the broader disruption patterns in mind. Six show up repeatedly.

Rapid commoditization. AI capabilities become widely available and every organization has access to broadly the same tools. Advantage from tooling alone shrinks toward zero, and the winners are the organizations that integrate AI most effectively into how they actually run.

Specialized capabilities emerge. Rather than general-purpose assistants dominating, capabilities tuned to specific domains become the important ones. Organizations with deep expertise in the AI of their own field pull ahead.

Workforce displacement. Automation reaches enough jobs that displacement becomes a societal issue rather than a team-level one. Organizations that handle it responsibly, by retraining displaced workers and creating new roles, come through it intact. Those that handle it badly face backlash.

Regulatory restrictions. Regulation tightens and AI can only be used in limited, approved ways. Competitive advantage moves to whoever navigates the rules most effectively. This is the wider version of Caleb's Future C.

Capability explosion. Capabilities expand faster than anyone expected and new applications keep appearing. The advantage goes to organizations that can innovate and adapt quickly rather than those with the best current plan.

Your industry disrupted. AI-capable competitors from other industries enter your market and the traditional business model stops working. Only organizations willing to adapt the model itself survive that one.

These are not equally likely, and nobody should treat them as predictions. They are all possible, which is exactly the point. A good strategy prepares for more than one of them.

Building Scenarios With a 2x2 Matrix

At team scale, three hand-written futures were enough. At organizational scale there is a more structured method, and it is worth knowing because it produces scenarios that are genuinely different from each other rather than three shades of the same guess.

Identify your key uncertainties. Ask which factors are both most uncertain and most consequential for your domain. Consider a customer service function. Its key uncertainties might be whether chatbot technology improves faster or slower than expected, whether customers accept AI-assisted service or demand human interaction, whether regulatory restrictions increase or stay light, and how aggressively competitors adopt AI. List five to ten such uncertainties for your own domain before narrowing.

Define your scenario axes. Take the two uncertainties that matter most and turn them into axes. In the customer service example, the horizontal axis is AI capability, running from low to high, and the vertical axis is the regulatory environment, running from permissive to restrictive. Two axes give you a 2x2 matrix and therefore four scenarios.

Describe each scenario as a world. Write each quadrant out until it feels like somewhere you could walk around in.

• Advanced AI, restrictive regulation. Capabilities advance rapidly and chatbots become very capable, but regulation limits when and how you may use them and certain decisions cannot be automated at all. The opportunity is to develop AI within the constraints, and advantage flows to whoever navigates regulation well.
• Advanced AI, permissive regulation. Capabilities advance rapidly and the rules stay light, so AI can be used broadly. The opportunity belongs to first movers who deploy aggressively and capture significant advantage before others react.
• Limited AI progress, restrictive regulation. Capability improves slowly and chatbots stay limited, while regulation restricts where AI can be applied. The opportunity is focused use of AI where it is genuinely most valuable, and human service roles remain important.
• Limited AI progress, permissive regulation. Capability improves slowly and the rules stay light. Adoption is gradual, and competitive advantage comes from operational excellence rather than from AI itself.

For each of the four, ask the same four questions Caleb asked of his three: what changes in this world, what opportunities emerge, what threats emerge, and what would we need to do to win.

Assess the implications. Then ask the question that makes the exercise pay. Which scenario is our current strategy optimized for, and if the world moves to a different one, does the strategy still work? If your strategy only functions in the advanced-and-permissive quadrant while the advanced-and-restrictive quadrant is equally likely, that is a vulnerability, and the right response is to adjust the strategy so it holds across more of the matrix.

What Makes a Strategy Robust

A robust strategy works across multiple scenarios. It is not tuned for one future and brittle everywhere else. Instead of betting entirely on deploying advanced chatbots, which only pays off in the advanced-and-permissive world, you invest in both AI capability and human skill development. That combination works when AI is unleashed, because you have both the tools and the people, and it also works when regulation restricts you or capability lags, because you lean harder on the people. Caleb reached the same conclusion at his own scale when he decided to reskill his analysts and pursue automation at the same time.

Robust strategies tend to share five characteristics.

• Flexibility. The strategy can adapt if the world shifts to a different scenario, because your investments have value in more than one of them.
• Reversibility. If something is not working you can change course without catastrophic loss, because you did not bet everything on a single approach.
• Optionality. You deliberately develop options for several scenarios and avoid committing fully to one path until the future becomes clearer.
• Early warnings. You identify the leading indicators that signal which scenario is emerging, and you accelerate investment in that direction when you see them. A leading indicator of the restrictive world, for instance, is regulatory bodies beginning to scrutinize AI use; on that signal you increase investment in compliance capability and domain expertise.
• A portfolio approach. Rather than one large initiative, you run several. Some are bets on high-impact scenarios, some are bets on more certain outcomes, and together they form a portfolio that produces value across futures.

Three Ways Scenario Planning Goes Wrong

The single-scenario bet. An organization commits everything to one future, investing enormously on the assumption that capability will advance rapidly and regulation will stay light. If the world lands somewhere else, the exposure is total. The alternative is to build strategies that hold across several scenarios and to avoid betting the whole organization on a single outcome.

Scenario paralysis. An organization spends so long building scenarios that it never acts, insisting it must understand all possible futures before investing. Meanwhile competitors move and it falls behind. Scenarios exist to inform strategy, not to postpone decisions. Invest despite the uncertainty, watch for the early warnings, and adjust as the future reveals itself.

Static scenarios. An organization builds its scenarios once and keeps using them for years. The world changes, key uncertainties resolve, new ones appear, and the old scenarios quietly become irrelevant. Refresh them on a schedule, exactly as Caleb refreshes his team plan each quarter, so your picture of the future keeps pace with the world.

Four Terms Worth Being Precise About

• Scenario planning: the strategic process of exploring multiple possible futures and preparing the organization for each of them.
• Uncertainty: a situation where multiple outcomes are possible and the probability of each is unknown.
• Leading indicators: signals that reveal which scenario is emerging, giving you time to adjust strategy before you are forced to.
• Robust strategy: a strategy that works across multiple possible futures rather than being optimized for a single one.

Practice and Reflection

Work these through with your own team and domain in front of you.

• Identify the three most important uncertainties about AI's future in your industry. Describe what each one could lead to, then decide which resulting scenarios are most important to prepare for.
• Build a 2x2 scenario matrix from two key uncertainties for your organization. Describe each of the four scenarios, then assess what each would mean for you.
• Evaluate your current AI strategy. Which scenario is it optimized for? How would it perform in the others? What single change would make it more robust?
• Identify the leading indicators that would signal which scenario is emerging. What specifically would you monitor to know which future is unfolding?

Then sit with three reflection questions. What is the most consequential uncertainty about AI's future for your industry? If you were genuinely preparing your organization for multiple AI futures, what would you do differently from what you are doing today? And what early warnings would tell you that one scenario is emerging rather than another?

Acting Under Uncertainty

Scenario planning is the bridge between uncertainty and action. It is what allows you to act strategically without knowing the future, which is the only condition under which anyone has ever had to act.

By this point in the certification you have learned to implement AI within your team, to scale it across an organization, to think strategically about it, and now to navigate the uncertainty ahead. The future with AI is not predetermined. It will be shaped by the choices organizations and leaders make now, and scenario planning is how you make those choices with some foresight rather than by reflex.

Go forward and use the frameworks. Keep learning, adapt as the world changes, and lead with integrity. Caleb never did get a clean answer to his analyst's question, because no such answer exists. What he got instead was a plan that worked under three different futures and a team that could see it. The future is uncertain. You are now equipped to navigate it.

Key Takeaways

• Think in tasks, not jobs. AI rarely replaces a whole job at once; it changes individual tasks. Breaking a role into tasks turns a frightening "will AI take my job" into a workable list of what to automate, augment, and protect.
• Map your team's task exposure first. List the real recurring tasks, estimate time spent on each, and rate exposure as High, Medium, or Low plus automate or augment. This shared picture is the foundation for every other decision.
• Build three plausible futures, not one prediction. Gradual augmentation, fast automation, and regulation slowing things down each demand different responses. Writing more than one stops you assuming the future you find most comfortable.
• Freed hours are a redeployment signal, not a layoff signal. Project how much capacity AI frees under each future and decide in advance where it goes, so the team grows into higher-value work instead of looking like surplus.
• Plan reskilling as named skills, scheduled time, and follow-ups. Vague encouragement to upskill fails. Concrete paths with protected learning time and check-in dates succeed.
• Redesign roles around what survives. Do not leave a hollowed-out role; rebuild it around judgment, relationships, and supervising AI, and share that picture early so people have something to move toward.
• Name your early signals before they appear. Decide in advance what would tell you a given future is arriving, so you adjust ahead of the curve instead of reacting late.
• Communicate honestly to reduce fear. Avoid both silence and over-reassurance. Separate task from job out loud, commit to what you control, and give people a path that turns fear into an active project.
• Revisit the plan every quarter. A scenario plan filed away goes stale and breeds false confidence. A short, recurring review keeps it alive and keeps your team protected.
• Robust beats optimized. A strategy tuned for one future is brittle. Flexibility, reversibility, optionality, early warnings, and a portfolio of bets are what let a strategy survive whichever future arrives.
• Use scenarios to decide, not to delay. Betting everything on one future, freezing until the future is clear, and never refreshing old scenarios are the three ways this discipline fails.